Da in letzter Zeit viele neue Modelle bzw. Updates erschienen sind, ist es interessant mal einige mit identischem Prompt zu testen.
Das ist natürlich kein echter Funktionsvergleich, zumal die Modelle teils sehr unterschiedliche Anweisungen benötigen.
A four-panel comedy comic strip, vertical layout. The style mixes futuristic cyberpunk elements with a mundane train station setting. Night time. Lots of lights.
Panel 1 (Top): A young slim woman with half long black hair, glasses, a big winter jacket, blue jeans and brown boots. She is sitting at a train station. It is winter and snow everywhere. A cute little dog is sitting beside her. She is frowning. A display behind her says „OUT OF SERVICE“. Her speech bubble says „Verdammt kalt hier…“ There is a half melted sad looking snowman standing next to her.
Panel 2 : The woman is standing up looking at an incoming train cramped up with people of different ethnicity. Text bubble (Woman thinking): „Alles voll.. „
Panel 3 : The woman is getting angry and walks by the train with the dog by her side. The train moves out of the station. Text bubble (Woman) : „So nicht!“
Panel 4 : The whole place is extremely crowded. The dog is holding up a sign „Ich mach Home Office“
Qwen 2512 Wuli-art/Qwen-Image-2512-Turbo-LoRA :
Z-Image-Turbo
Z-Image-Turbo with LoRa
Flux.2 Dev 8steps
Flux.2 Dev 8steps mit Referenzbild (hier muss man den Zusatz ‚Keep the face and the clothes‘ im Prompt einfügen.
